Convolutional Neural Networks

Project: Write an Algorithm for a Dog Identification App


In this notebook, some template code has already been provided for you, and you will need to implement additional functionality to successfully complete this project. You will not need to modify the included code beyond what is requested. Sections that begin with '(IMPLEMENTATION)' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section, and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!

Note: Once you have completed all of the code implementations, you need to finalize your work by exporting the Jupyter Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.

In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.

Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. Markdown cells can be edited by double-clicking the cell to enter edit mode.

The rubric contains optional "Stand Out Suggestions" for enhancing the project beyond the minimum requirements. If you decide to pursue the "Stand Out Suggestions", you should include the code in this Jupyter notebook.


Why We're Here

In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project (... but we expect that each student's algorithm will behave differently!).

Sample Dog Output

In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!

The Road Ahead

We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.

  • Step 0: Import Datasets
  • Step 1: Detect Humans
  • Step 2: Detect Dogs
  • Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
  • Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)
  • Step 5: Write your Algorithm
  • Step 6: Test Your Algorithm

Step 0: Import Datasets

Make sure that you've downloaded the required human and dog datasets:

  • Download the dog dataset. Unzip the folder and place it in this project's home directory, at the location /dogImages.

  • Download the human dataset. Unzip the folder and place it in the home diretcory, at location /lfw.

Note: If you are using a Windows machine, you are encouraged to use 7zip to extract the folder.

In the code cell below, we save the file paths for both the human (LFW) dataset and dog dataset in the numpy arrays human_files and dog_files.

In [96]:
import numpy as np
from glob import glob

# load filenames for human and dog images
human_files = np.array(glob("lfw/*/*"))
dog_files = np.array(glob("dogImages/*/*/*"))

# print number of images in each dataset
print('There are %d total human images.' % len(human_files))
print('There are %d total dog images.' % len(dog_files))
There are 13233 total human images.
There are 8350 total dog images.

Step 1: Detect Humans

In this section, we use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images.

OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory. In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.

In [97]:
import cv2                
import matplotlib.pyplot as plt                        
%matplotlib inline                               

# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

# load color (BGR) image
img = cv2.imread(human_files[0])
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# find faces in image
faces = face_cascade.detectMultiScale(gray)

# print number of faces detected in the image
print('Number of faces detected:', len(faces))

# get bounding box for each detected face
for (x,y,w,h) in faces:
    # add bounding box to color image
    cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
    
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Number of faces detected: 1

Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.

In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.

Write a Human Face Detector

We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.

In [98]:
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
    img = cv2.imread(img_path)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(gray)
    return len(faces) > 0

(IMPLEMENTATION) Assess the Human Face Detector

Question 1: Use the code cell below to test the performance of the face_detector function.

  • What percentage of the first 100 images in human_files have a detected human face?
  • What percentage of the first 100 images in dog_files have a detected human face?

Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.

Answer: (You can print out your results and/or write your percentages in this cell)

In [99]:
from tqdm import tqdm

human_files_short = human_files[:100]
dog_files_short = dog_files[:100]

#-#-# Do NOT modify the code above this line. #-#-#

## TODO: Test the performance of the face_detector algorithm 
## on the images in human_files_short and dog_files_short.
print("Accuracy on the human dataset is {}%".format(sum([face_detector(img) for img in human_files_short])))
print("Accuracy on the dog dataset is {}%".format(sum([face_detector(img) for img in dog_files_short])))
Accuracy on the human dataset is 99%
Accuracy on the dog dataset is 13%

We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [100]:
import face_recognition
In [101]:
image = face_recognition.load_image_file(human_files_short[5])
face_locations = face_recognition.face_locations(image)
for (x1, y1, x2, y2) in face_locations:
    cv2.rectangle(image,(x1,y1),(x2,y2),(255,0,0),2)
plt.imshow(image)
plt.show()
In [102]:
# Let us try this on a dog image
image = face_recognition.load_image_file('dogImages/test/002.Afghan_hound/Afghan_hound_00151.jpg')
face_locations = face_recognition.face_locations(image)
for (x1, y1, x2, y2) in face_locations:
    cv2.rectangle(image,(x1,y1),(x2,y2),(255,0,0),2)
plt.imshow(image)
plt.show()
In [103]:
def face_detector_using_face_recognition(img_path, dog=False):
    image = face_recognition.load_image_file(img_path)
    face_locations = face_recognition.face_locations(image)
    if dog and len(face_locations)>0:
        print('\t{}'.format(img_path))
    return len(face_locations) > 0
In [104]:
print("face_recognition's accuracy on the human dataset is {}%".format(sum([face_detector_using_face_recognition(img) for img in human_files_short])))
print("face_recognition's accuracy on the dog dataset is {}%".format(sum([face_detector_using_face_recognition(img, True) for img in dog_files_short])))
face_recognition's accuracy on the human dataset is 99%
	dogImages/test/015.Basset_hound/Basset_hound_01097.jpg
	dogImages/test/048.Chihuahua/Chihuahua_03460.jpg
	dogImages/test/048.Chihuahua/Chihuahua_03405.jpg
	dogImages/test/058.Dandie_dinmont_terrier/Dandie_dinmont_terrier_04152.jpg
	dogImages/test/058.Dandie_dinmont_terrier/Dandie_dinmont_terrier_04134.jpg
	dogImages/test/002.Afghan_hound/Afghan_hound_00151.jpg
	dogImages/test/076.Golden_retriever/Golden_retriever_05221.jpg
	dogImages/test/076.Golden_retriever/Golden_retriever_05240.jpg
	dogImages/test/076.Golden_retriever/Golden_retriever_05258.jpg
face_recognition's accuracy on the dog dataset is 9%
In [105]:
def show_faces_in_dog_images(img_path):
    image = face_recognition.load_image_file(img_path)
    face_locations = face_recognition.face_locations(image)
    for (x1, y1, x2, y2) in face_locations:
        cv2.rectangle(image,(x1,y1),(x2,y2),(255,0,0),2)
    plt.imshow(image)
    plt.show()
In [106]:
dog_images_with_faces = [
    'dogImages/test/015.Basset_hound/Basset_hound_01097.jpg',
    'dogImages/test/048.Chihuahua/Chihuahua_03460.jpg',
    'dogImages/test/048.Chihuahua/Chihuahua_03405.jpg',
    'dogImages/test/058.Dandie_dinmont_terrier/Dandie_dinmont_terrier_04152.jpg',
    'dogImages/test/058.Dandie_dinmont_terrier/Dandie_dinmont_terrier_04134.jpg',
    'dogImages/test/002.Afghan_hound/Afghan_hound_00151.jpg',
    'dogImages/test/076.Golden_retriever/Golden_retriever_05221.jpg',
    'dogImages/test/076.Golden_retriever/Golden_retriever_05240.jpg',
    'dogImages/test/076.Golden_retriever/Golden_retriever_05258.jpg'
]
In [107]:
for image in dog_images_with_faces:
    show_faces_in_dog_images(image)

Step 2: Detect Dogs

In this section, we use a pre-trained model to detect dogs in images.

Obtain Pre-trained VGG-16 Model

The code cell below downloads the VGG-16 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories.

In [108]:
import torch
import torchvision.models as models

# define VGG16 model
vgg16 = models.vgg16(pretrained=True)

# check if CUDA is available
use_cuda = torch.cuda.is_available()

# move model to GPU if CUDA is available
if use_cuda:
    vgg16.cuda()
In [109]:
# Unit test with the model
from PIL import Image
import torchvision.transforms as transforms


normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                 std=[0.229, 0.224, 0.225])

transform = transforms.Compose([
    transforms.RandomResizedCrop(224),
    transforms.RandomHorizontalFlip(),
    transforms.ToTensor(),
    normalize
])

Given an image, this pre-trained VGG-16 model returns a prediction (derived from the 1000 possible categories in ImageNet) for the object that is contained in the image.

(IMPLEMENTATION) Making Predictions with a Pre-trained Model

In the next code cell, you will write a function that accepts a path to an image (such as 'dogImages/train/001.Affenpinscher/Affenpinscher_00001.jpg') as input and returns the index corresponding to the ImageNet class that is predicted by the pre-trained VGG-16 model. The output should always be an integer between 0 and 999, inclusive.

Before writing the function, make sure that you take the time to learn how to appropriately pre-process tensors for pre-trained models in the PyTorch documentation.

In [110]:
def VGG16_predict(img_path):
    '''
    Use pre-trained VGG-16 model to obtain index corresponding to 
    predicted ImageNet class for image at specified path
    
    Args:
        img_path: path to an image
        
    Returns:
        Index corresponding to VGG-16 model's prediction
    '''
    
    ## TODO: Complete the function.
    ## Load and pre-process an image from the given img_path
    ## Return the *index* of the predicted class for that image
    img = Image.open(img_path)
    img = transform(img)
    c, h, w = img.size()
    img = img.view((1, c, h, w))
    if use_cuda:
        img = img.cuda()
    output = vgg16(img)
    _, pred = torch.max(output.data, 1)
    return pred # predicted class index

(IMPLEMENTATION) Write a Dog Detector

While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained VGG-16 model, we need only check if the pre-trained model predicts an index between 151 and 268 (inclusive).

Use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).

In [111]:
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
    ## TODO: Complete the function.
    pred = VGG16_predict(img_path)
    return True if pred >= 151 and pred <= 268 else False # true/false

(IMPLEMENTATION) Assess the Dog Detector

Question 2: Use the code cell below to test the performance of your dog_detector function.

  • What percentage of the images in human_files_short have a detected dog?
  • What percentage of the images in dog_files_short have a detected dog?

Answer: The dog_detector detected 2% dogs in the truncated human dataset while it detected 98% of all dogs in the truncated dog dataset.

In [112]:
### TODO: Test the performance of the dog_detector function
### on the images in human_files_short and dog_files_short.
print("Accuracy on the human dataset is {}%".format(sum([dog_detector(img) for img in human_files_short])))
print("Accuracy on the dog dataset is {}%".format(sum([dog_detector(img) for img in dog_files_short])))
Accuracy on the human dataset is 2%
Accuracy on the dog dataset is 98%

We suggest VGG-16 as a potential network to detect dog images in your algorithm, but you are free to explore other pre-trained networks (such as Inception-v3, ResNet-50, etc). Please use the code cell below to test other pre-trained PyTorch models. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [113]:
### (Optional) 
### TODO: Report the performance of another pre-trained network.
### Feel free to use as many code cells as needed.
densenet201 = models.densenet201(pretrained=True)

if use_cuda:
    densenet201.cuda()

def densenet201_predict(img_path):
    img = Image.open(img_path)
    img = transform(img)
    c, h, w = img.size()
    img = img.view((1, c, h, w))
    if use_cuda:
        img = img.cuda()
    output = densenet201(img)
    _, pred = torch.max(output.data, 1)
    return pred # predicted class index
In [114]:
def dog_detector_densenet201(img_path):
    ## TODO: Complete the function.
    pred = densenet201_predict(img_path)
    return True if pred >= 151 and pred <= 268 else False # true/false
In [115]:
print("Accuracy on the human dataset is {}%".format(sum([dog_detector_densenet201(img) for img in human_files_short])))
print("Accuracy on the dog dataset is {}%".format(sum([dog_detector_densenet201(img) for img in dog_files_short])))
Accuracy on the human dataset is 0%
Accuracy on the dog dataset is 0%
In [116]:
sq = models.squeezenet1_1(pretrained=True)

if use_cuda:
    sq.cuda()

def sq_predict(img_path):
    img = Image.open(img_path)
    img = transform(img)
    c, h, w = img.size()
    img = img.view((1, c, h, w))
    if use_cuda:
        img = img.cuda()
    output = sq(img)
    _, pred = torch.max(output.data, 1)
    return pred # predicted class index

def dog_detector_sq(img_path):
    ## TODO: Complete the function.
    pred = sq_predict(img_path)
    return True if pred >= 151 and pred <= 268 else False # true/false

print("Accuracy on the human dataset is {}%".format(sum([dog_detector_sq(img) for img in human_files_short])))
print("Accuracy on the dog dataset is {}%".format(sum([dog_detector_sq(img) for img in dog_files_short])))
Accuracy on the human dataset is 8%
Accuracy on the dog dataset is 92%

Step 3: Create a CNN to Classify Dog Breeds (from Scratch)

Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 10%. In Step 4 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.

We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have trouble distinguishing between a Brittany and a Welsh Springer Spaniel.

Brittany Welsh Springer Spaniel

It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).

Curly-Coated Retriever American Water Spaniel

Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.

Yellow Labrador Chocolate Labrador Black Labrador

We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.

Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dogImages/train, dogImages/valid, and dogImages/test, respectively). You may find this documentation on custom datasets to be a useful resource. If you are interested in augmenting your training and/or validation data, check out the wide variety of transforms!

In [117]:
import os
from torchvision import datasets
from torch.utils.data import DataLoader as dl

### TODO: Write data loaders for training, validation, and test sets
## Specify appropriate transforms, and batch_sizes


trfm = transforms.Compose([
    transforms.RandomResizedCrop(224),
    transforms.RandomRotation(10),
    transforms.RandomHorizontalFlip(),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225])
])

train_data = datasets.ImageFolder('dogImages/train', transform=trfm)
valid_data = datasets.ImageFolder('dogImages/valid', transform=trfm)
test_data = datasets.ImageFolder('dogImages/test', transform=trfm)

train_loader = dl(train_data,
                  batch_size=16,
                  shuffle=True,
                  num_workers=4,
                 )

valid_loader = dl(valid_data,
                  batch_size=16,
                  shuffle=True,
                  num_workers=4)

test_loader = dl(test_data,
                  batch_size=16,
                  shuffle=True,
                  num_workers=4)

loaders_scratch = {
    'train' : train_loader,
    'valid' : valid_loader,
    'test'  : test_loader
}

Question 3: Describe your chosen procedure for preprocessing the data.

  • How does your code resize the images (by cropping, stretching, etc)? What size did you pick for the input tensor, and why?
  • Did you decide to augment the dataset? If so, how (through translations, flips, rotations, etc)? If not, why not?

Answer:

  • I use RandomResizedCrop to resize the image. This function scales the image with a certain randomness and then resizes the image to the demanded size. The size I picked for the input tensor was 224x224. I checked the images and most of the images were larger than these dimensions and these dimensions also did better than 32x32 which I chose earlier. Have a larger input tensor size, gave the network a better chance to capture detailed features which is needed in the case of distinguishing between dogs of different breeds.
  • I did decide to augment the dataset using random rotations and random horizontal flips. This afforded the network the opportunity to learn more features which were location and angle invariant and perform better on the test set.

(IMPLEMENTATION) Model Architecture

Create a CNN to classify dog breed. Use the template in the code cell below.

In [118]:
import torch.nn as nn
import torch.nn.functional as F

# define the CNN architecture
class Net(nn.Module):
    ### TODO: choose an architecture, and complete the class
    def __init__(self):
        super(Net, self).__init__()
        ## Define layers of a CNN
        # convolutional layer
        self.conv1 = nn.Conv2d(3, 16, 3, padding=1)
        # max pooling layer
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(16, 32, 3, padding=1)
        self.conv3 = nn.Conv2d(32, 64, 3, padding=1)
        self.fc1 = nn.Linear(64 * 28 * 28, 1024)
        self.fc2 = nn.Linear(1024, 512)
        self.fc3 = nn.Linear(512, 133)
    
    def forward(self, x):
        ## Define forward behavior
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = self.pool(F.relu(self.conv3(x)))
        x = x.view(-1, 64 * 28 * 28)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)
        return x

#-#-# You so NOT have to modify the code below this line. #-#-#

# instantiate the CNN
model_scratch = Net()
# move tensors to GPU if CUDA is available
if use_cuda:
    model_scratch.cuda()
    
# Initialize model weights
def init_weights(m):
    if type(m) == nn.Linear:
        torch.nn.init.xavier_uniform_(m.weight)
        m.bias.data.fill_(0.01)

model_scratch = model_scratch.apply(init_weights)

Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step.

Answer: Since the input tensor was quite large, the network had more data to work with. I had the luxury of GPUs and data, so I chose to add a 3 Convolution layers and 3 fully connected layers. While the network trained faster with lesser conv layers, the accuracy was suffering. The network arch I choose was also inspired by earlier in class exercies which also had similar structures. This particular structure seems to be good at extracting high level features. The output after 3 conv layers and consequent pooling gives us a 64 28 28 tensor. I didn't want to dramatically bring them down in one step to 133 classification classes, so I decided to add a 3 fully connected layers and gradually reduce the number of features to 133. I also found that initializing the weights for the fully connected layers helped a lot. I experimented with and without initialization and noticed that the validation loss decreased dramatically when the weights were initialized using xavier_uniform_ initliaziation function. Hence, I wrote a init_weights function to initialize the weights of the linear layers.

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_scratch, and the optimizer as optimizer_scratch below.

In [119]:
import torch.optim as optim

# specify loss function
criterion_scratch = nn.CrossEntropyLoss()

# specify optimizer
optimizer_scratch = optim.SGD(model_scratch.parameters(), lr = 0.01, momentum=0.9)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_scratch.pt'.

In [120]:
def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
    """returns trained model"""
    # initialize tracker for minimum validation loss
    valid_loss_min = np.Inf 
    
    for epoch in range(1, n_epochs+1):
        # initialize variables to monitor training and validation loss
        train_loss = 0.0
        valid_loss = 0.0
        
        ###################
        # train the model #
        ###################
        for batch_idx, (data, target) in enumerate(loaders['train']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            ## find the loss and update the model parameters accordingly
            ## record the average training loss, using something like
            ## train_loss = train_loss + ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            optimizer.zero_grad()
            # forward pass: compute predicted outputs by passing inputs to the model
            output = model(data)
            # calculate the batch loss
            loss = criterion(output, target)
            # backward pass: compute gradient of the loss with respect to model parameters
            loss.backward()
            # perform a single optimization step (parameter update)
            optimizer.step()
            # update training loss
            train_loss += ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
        ######################    
        # validate the model #
        ######################
        for batch_idx, (data, target) in enumerate(loaders['valid']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            ## update the average validation loss
            # forward pass: compute predicted outputs by passing inputs to the model
            output = model(data)
            # calculate the batch loss
            loss = criterion(output, target)
            # update average validation loss 
            valid_loss += ((1 / (batch_idx + 1)) * (loss.data - valid_loss))

            
        # print training/validation statistics 
        print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format(
            epoch, 
            train_loss,
            valid_loss
            ))
        
        ## TODO: save the model if validation loss has decreased
        if valid_loss < valid_loss_min:
            print('Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...'.format(
                valid_loss_min,
                valid_loss))
            valid_loss_min = valid_loss
            # save model
            torch.save(model.state_dict(), save_path)

            
            
    # return trained model
    return model


# train the model
model_scratch = train(100, loaders_scratch, model_scratch, optimizer_scratch, 
                      criterion_scratch, use_cuda, 'model_scratch.pt')

# load the model that got the best validation accuracy
model_scratch.load_state_dict(torch.load('model_scratch.pt'))
Epoch: 1 	Training Loss: 4.827129 	Validation Loss: 4.712269
Validation loss decreased (inf --> 4.712269).  Saving model ...
Epoch: 2 	Training Loss: 4.690363 	Validation Loss: 4.632637
Validation loss decreased (4.712269 --> 4.632637).  Saving model ...
Epoch: 3 	Training Loss: 4.609834 	Validation Loss: 4.586682
Validation loss decreased (4.632637 --> 4.586682).  Saving model ...
Epoch: 4 	Training Loss: 4.541701 	Validation Loss: 4.530297
Validation loss decreased (4.586682 --> 4.530297).  Saving model ...
Epoch: 5 	Training Loss: 4.519663 	Validation Loss: 4.538045
Epoch: 6 	Training Loss: 4.450006 	Validation Loss: 4.468002
Validation loss decreased (4.530297 --> 4.468002).  Saving model ...
Epoch: 7 	Training Loss: 4.423037 	Validation Loss: 4.425876
Validation loss decreased (4.468002 --> 4.425876).  Saving model ...
Epoch: 8 	Training Loss: 4.375903 	Validation Loss: 4.365127
Validation loss decreased (4.425876 --> 4.365127).  Saving model ...
Epoch: 9 	Training Loss: 4.324154 	Validation Loss: 4.293947
Validation loss decreased (4.365127 --> 4.293947).  Saving model ...
Epoch: 10 	Training Loss: 4.303766 	Validation Loss: 4.414248
Epoch: 11 	Training Loss: 4.254137 	Validation Loss: 4.349981
Epoch: 12 	Training Loss: 4.231278 	Validation Loss: 4.239252
Validation loss decreased (4.293947 --> 4.239252).  Saving model ...
Epoch: 13 	Training Loss: 4.203918 	Validation Loss: 4.349658
Epoch: 14 	Training Loss: 4.161320 	Validation Loss: 4.354726
Epoch: 15 	Training Loss: 4.100708 	Validation Loss: 4.204844
Validation loss decreased (4.239252 --> 4.204844).  Saving model ...
Epoch: 16 	Training Loss: 4.071532 	Validation Loss: 4.291914
Epoch: 17 	Training Loss: 4.053299 	Validation Loss: 4.215440
Epoch: 18 	Training Loss: 4.016264 	Validation Loss: 4.177547
Validation loss decreased (4.204844 --> 4.177547).  Saving model ...
Epoch: 19 	Training Loss: 3.990551 	Validation Loss: 4.115688
Validation loss decreased (4.177547 --> 4.115688).  Saving model ...
Epoch: 20 	Training Loss: 3.971789 	Validation Loss: 4.095636
Validation loss decreased (4.115688 --> 4.095636).  Saving model ...
Epoch: 21 	Training Loss: 3.915852 	Validation Loss: 4.111567
Epoch: 22 	Training Loss: 3.931022 	Validation Loss: 4.117650
Epoch: 23 	Training Loss: 3.890080 	Validation Loss: 4.197191
Epoch: 24 	Training Loss: 3.865591 	Validation Loss: 4.042549
Validation loss decreased (4.095636 --> 4.042549).  Saving model ...
Epoch: 25 	Training Loss: 3.841220 	Validation Loss: 3.990913
Validation loss decreased (4.042549 --> 3.990913).  Saving model ...
Epoch: 26 	Training Loss: 3.816364 	Validation Loss: 4.203636
Epoch: 27 	Training Loss: 3.773868 	Validation Loss: 4.070373
Epoch: 28 	Training Loss: 3.764486 	Validation Loss: 4.010434
Epoch: 29 	Training Loss: 3.760958 	Validation Loss: 4.096935
Epoch: 30 	Training Loss: 3.768084 	Validation Loss: 4.102507
Epoch: 31 	Training Loss: 3.741421 	Validation Loss: 4.064619
Epoch: 32 	Training Loss: 3.728990 	Validation Loss: 4.065399
Epoch: 33 	Training Loss: 3.698824 	Validation Loss: 4.021433
Epoch: 34 	Training Loss: 3.699548 	Validation Loss: 4.033219
Epoch: 35 	Training Loss: 3.661783 	Validation Loss: 4.090598
Epoch: 36 	Training Loss: 3.637377 	Validation Loss: 4.041481
Epoch: 37 	Training Loss: 3.638158 	Validation Loss: 4.083047
Epoch: 38 	Training Loss: 3.645388 	Validation Loss: 3.940708
Validation loss decreased (3.990913 --> 3.940708).  Saving model ...
Epoch: 39 	Training Loss: 3.604541 	Validation Loss: 4.100962
Epoch: 40 	Training Loss: 3.589895 	Validation Loss: 4.066535
Epoch: 41 	Training Loss: 3.589450 	Validation Loss: 4.078755
Epoch: 42 	Training Loss: 3.592707 	Validation Loss: 3.988340
Epoch: 43 	Training Loss: 3.579027 	Validation Loss: 4.005534
Epoch: 44 	Training Loss: 3.577658 	Validation Loss: 4.116406
Epoch: 45 	Training Loss: 3.563737 	Validation Loss: 3.975332
Epoch: 46 	Training Loss: 3.504154 	Validation Loss: 4.064783
Epoch: 47 	Training Loss: 3.511627 	Validation Loss: 4.011067
Epoch: 48 	Training Loss: 3.501081 	Validation Loss: 4.047154
Epoch: 49 	Training Loss: 3.524997 	Validation Loss: 4.267590
Epoch: 50 	Training Loss: 3.518666 	Validation Loss: 4.147994
Epoch: 51 	Training Loss: 3.478348 	Validation Loss: 3.928013
Validation loss decreased (3.940708 --> 3.928013).  Saving model ...
Epoch: 52 	Training Loss: 3.468640 	Validation Loss: 4.104277
Epoch: 53 	Training Loss: 3.496659 	Validation Loss: 4.171969
Epoch: 54 	Training Loss: 3.456491 	Validation Loss: 4.134822
Epoch: 55 	Training Loss: 3.448495 	Validation Loss: 4.055060
Epoch: 56 	Training Loss: 3.446239 	Validation Loss: 4.091902
Epoch: 57 	Training Loss: 3.441341 	Validation Loss: 4.011936
Epoch: 58 	Training Loss: 3.448008 	Validation Loss: 3.903837
Validation loss decreased (3.928013 --> 3.903837).  Saving model ...
Epoch: 59 	Training Loss: 3.421675 	Validation Loss: 3.973899
Epoch: 60 	Training Loss: 3.414291 	Validation Loss: 4.005034
Epoch: 61 	Training Loss: 3.472571 	Validation Loss: 4.083932
Epoch: 62 	Training Loss: 3.431108 	Validation Loss: 4.286971
Epoch: 63 	Training Loss: 3.422911 	Validation Loss: 4.296069
Epoch: 64 	Training Loss: 3.443967 	Validation Loss: 4.059336
Epoch: 65 	Training Loss: 3.388727 	Validation Loss: 4.149616
Epoch: 66 	Training Loss: 3.448043 	Validation Loss: 3.940369
Epoch: 67 	Training Loss: 3.411679 	Validation Loss: 4.245986
Epoch: 68 	Training Loss: 3.387933 	Validation Loss: 4.145841
Epoch: 69 	Training Loss: 3.419027 	Validation Loss: 4.394275
Epoch: 70 	Training Loss: 3.403261 	Validation Loss: 4.138064
Epoch: 71 	Training Loss: 3.451111 	Validation Loss: 4.076566
Epoch: 72 	Training Loss: 3.399917 	Validation Loss: 4.180845
Epoch: 73 	Training Loss: 3.383409 	Validation Loss: 4.136524
Epoch: 74 	Training Loss: 3.384444 	Validation Loss: 4.312338
Epoch: 75 	Training Loss: 3.425508 	Validation Loss: 4.024800
Epoch: 76 	Training Loss: 3.396508 	Validation Loss: 4.175879
Epoch: 77 	Training Loss: 3.375806 	Validation Loss: 4.143928
Epoch: 78 	Training Loss: 3.362357 	Validation Loss: 4.151093
Epoch: 79 	Training Loss: 3.391891 	Validation Loss: 4.110629
Epoch: 80 	Training Loss: 3.381522 	Validation Loss: 4.248757
Epoch: 81 	Training Loss: 3.355518 	Validation Loss: 4.075140
Epoch: 82 	Training Loss: 3.345594 	Validation Loss: 4.275232
Epoch: 83 	Training Loss: 3.380921 	Validation Loss: 4.081524
Epoch: 84 	Training Loss: 3.388050 	Validation Loss: 4.130834
Epoch: 85 	Training Loss: 3.336048 	Validation Loss: 4.249877
Epoch: 86 	Training Loss: 3.376682 	Validation Loss: 4.080374
Epoch: 87 	Training Loss: 3.369478 	Validation Loss: 4.086947
Epoch: 88 	Training Loss: 3.368992 	Validation Loss: 4.265532
Epoch: 89 	Training Loss: 3.315338 	Validation Loss: 4.129161
Epoch: 90 	Training Loss: 3.349287 	Validation Loss: 4.132030
Epoch: 91 	Training Loss: 3.291341 	Validation Loss: 4.055668
Epoch: 92 	Training Loss: 3.281848 	Validation Loss: 4.169878
Epoch: 93 	Training Loss: 3.317730 	Validation Loss: 4.246885
Epoch: 94 	Training Loss: 3.314806 	Validation Loss: 4.153877
Epoch: 95 	Training Loss: 3.295206 	Validation Loss: 4.229290
Epoch: 96 	Training Loss: 3.326927 	Validation Loss: 4.070469
Epoch: 97 	Training Loss: 3.270913 	Validation Loss: 4.205180
Epoch: 98 	Training Loss: 3.328300 	Validation Loss: 4.092924
Epoch: 99 	Training Loss: 3.322307 	Validation Loss: 4.244794
Epoch: 100 	Training Loss: 3.330589 	Validation Loss: 4.163348

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 10%.

In [121]:
def test(loaders, model, criterion, use_cuda):

    # monitor test loss and accuracy
    test_loss = 0.
    correct = 0.
    total = 0.

    for batch_idx, (data, target) in enumerate(loaders['test']):
        # move to GPU
        if use_cuda:
            data, target = data.cuda(), target.cuda()
        # forward pass: compute predicted outputs by passing inputs to the model
        output = model(data)
        # calculate the loss
        loss = criterion(output, target)
        # update average test loss 
        test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
        # convert output probabilities to predicted class
        pred = output.data.max(1, keepdim=True)[1]
        # compare predictions to true label
        correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
        total += data.size(0)
            
    print('Test Loss: {:.6f}\n'.format(test_loss))

    print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
        100. * correct / total, correct, total))

# call test function    
test(loaders_scratch, model_scratch, criterion_scratch, use_cuda)
Test Loss: 3.935390


Test Accuracy: 12% (103/836)

Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)

You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dogImages/train, dogImages/valid, and dogImages/test, respectively).

If you like, you are welcome to use the same data loaders from the previous step, when you created a CNN from scratch.

In [122]:
## TODO: Specify data loaders
loaders_transfer = loaders_scratch

(IMPLEMENTATION) Model Architecture

Use transfer learning to create a CNN to classify dog breed. Use the code cell below, and save your initialized model as the variable model_transfer.

In [123]:
model_import = models.vgg16(pretrained=True)
print(model_import)
VGG(
  (features): Sequential(
    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU(inplace)
    (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU(inplace)
    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (6): ReLU(inplace)
    (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (8): ReLU(inplace)
    (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (11): ReLU(inplace)
    (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (13): ReLU(inplace)
    (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (15): ReLU(inplace)
    (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (18): ReLU(inplace)
    (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (20): ReLU(inplace)
    (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (22): ReLU(inplace)
    (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (25): ReLU(inplace)
    (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (27): ReLU(inplace)
    (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (29): ReLU(inplace)
    (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  )
  (classifier): Sequential(
    (0): Linear(in_features=25088, out_features=4096, bias=True)
    (1): ReLU(inplace)
    (2): Dropout(p=0.5)
    (3): Linear(in_features=4096, out_features=4096, bias=True)
    (4): ReLU(inplace)
    (5): Dropout(p=0.5)
    (6): Linear(in_features=4096, out_features=1000, bias=True)
  )
)
In [124]:
# freeze weights
for param in model_import.parameters():
    param.require_grad = False
    
model_import.classifier[6] = nn.Linear(4096, 133)
print(model_import)
VGG(
  (features): Sequential(
    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (1): ReLU(inplace)
    (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (3): ReLU(inplace)
    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (6): ReLU(inplace)
    (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (8): ReLU(inplace)
    (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (11): ReLU(inplace)
    (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (13): ReLU(inplace)
    (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (15): ReLU(inplace)
    (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (18): ReLU(inplace)
    (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (20): ReLU(inplace)
    (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (22): ReLU(inplace)
    (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
    (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (25): ReLU(inplace)
    (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (27): ReLU(inplace)
    (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
    (29): ReLU(inplace)
    (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  )
  (classifier): Sequential(
    (0): Linear(in_features=25088, out_features=4096, bias=True)
    (1): ReLU(inplace)
    (2): Dropout(p=0.5)
    (3): Linear(in_features=4096, out_features=4096, bias=True)
    (4): ReLU(inplace)
    (5): Dropout(p=0.5)
    (6): Linear(in_features=4096, out_features=133, bias=True)
  )
)
In [125]:
torch.cuda.empty_cache()

model_transfer = model_import

if use_cuda: 
    model_transfer.cuda()

Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.

Answer: The network I created in step 4 was clearly not the best for this task. I felt like I needed a network which was deeper and capable of understanding more complex features. I experimented with resnet and squeezenet, and fell back to vgg16 and felt vgg16 is a pretty straight forward netowrk for this task. It is a great feature extractor and comfortably performs better than the 60% threshold. All I did was change the last fully connected layer to have 133 out features instead of 1000. Since vgg16 is already trained to classify dog breeds via ImageNet, it didn't take much to get the network to classify between 133 dog breeds.

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_transfer, and the optimizer as optimizer_transfer below.

In [126]:
criterion_transfer = nn.CrossEntropyLoss()
optimizer_transfer = optim.SGD(model_transfer.classifier.parameters(), lr=0.001)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_transfer.pt'.

In [127]:
# train the model
model_transfer = train(100, loaders_transfer, model_transfer, optimizer_transfer, criterion_transfer, use_cuda, 'model_transfer.pt')

# load the model that got the best validation accuracy (uncomment the line below)
model_transfer.load_state_dict(torch.load('model_transfer.pt'))
Epoch: 1 	Training Loss: 4.111222 	Validation Loss: 3.175933
Validation loss decreased (inf --> 3.175933).  Saving model ...
Epoch: 2 	Training Loss: 2.465518 	Validation Loss: 2.019856
Validation loss decreased (3.175933 --> 2.019856).  Saving model ...
Epoch: 3 	Training Loss: 1.761010 	Validation Loss: 1.609839
Validation loss decreased (2.019856 --> 1.609839).  Saving model ...
Epoch: 4 	Training Loss: 1.531374 	Validation Loss: 1.554820
Validation loss decreased (1.609839 --> 1.554820).  Saving model ...
Epoch: 5 	Training Loss: 1.392378 	Validation Loss: 1.382891
Validation loss decreased (1.554820 --> 1.382891).  Saving model ...
Epoch: 6 	Training Loss: 1.297469 	Validation Loss: 1.388928
Epoch: 7 	Training Loss: 1.223953 	Validation Loss: 1.376631
Validation loss decreased (1.382891 --> 1.376631).  Saving model ...
Epoch: 8 	Training Loss: 1.170249 	Validation Loss: 1.327944
Validation loss decreased (1.376631 --> 1.327944).  Saving model ...
Epoch: 9 	Training Loss: 1.137475 	Validation Loss: 1.204336
Validation loss decreased (1.327944 --> 1.204336).  Saving model ...
Epoch: 10 	Training Loss: 1.126811 	Validation Loss: 1.274288
Epoch: 11 	Training Loss: 1.111900 	Validation Loss: 1.208883
Epoch: 12 	Training Loss: 1.048806 	Validation Loss: 1.170147
Validation loss decreased (1.204336 --> 1.170147).  Saving model ...
Epoch: 13 	Training Loss: 1.037443 	Validation Loss: 1.129490
Validation loss decreased (1.170147 --> 1.129490).  Saving model ...
Epoch: 14 	Training Loss: 1.013504 	Validation Loss: 1.161186
Epoch: 15 	Training Loss: 1.020026 	Validation Loss: 1.217830
Epoch: 16 	Training Loss: 1.018075 	Validation Loss: 1.201443
Epoch: 17 	Training Loss: 0.960753 	Validation Loss: 1.123429
Validation loss decreased (1.129490 --> 1.123429).  Saving model ...
Epoch: 18 	Training Loss: 0.963017 	Validation Loss: 1.221394
Epoch: 19 	Training Loss: 0.943939 	Validation Loss: 1.205101
Epoch: 20 	Training Loss: 0.927068 	Validation Loss: 1.167333
Epoch: 21 	Training Loss: 0.951710 	Validation Loss: 1.184575
Epoch: 22 	Training Loss: 0.952654 	Validation Loss: 1.099233
Validation loss decreased (1.123429 --> 1.099233).  Saving model ...
Epoch: 23 	Training Loss: 0.923002 	Validation Loss: 1.175027
Epoch: 24 	Training Loss: 0.921284 	Validation Loss: 1.043582
Validation loss decreased (1.099233 --> 1.043582).  Saving model ...
Epoch: 25 	Training Loss: 0.901825 	Validation Loss: 1.142082
Epoch: 26 	Training Loss: 0.894199 	Validation Loss: 1.094130
Epoch: 27 	Training Loss: 0.923175 	Validation Loss: 1.092530
Epoch: 28 	Training Loss: 0.884691 	Validation Loss: 1.150740
Epoch: 29 	Training Loss: 0.887199 	Validation Loss: 1.128042
Epoch: 30 	Training Loss: 0.852261 	Validation Loss: 1.050473
Epoch: 31 	Training Loss: 0.860184 	Validation Loss: 0.961224
Validation loss decreased (1.043582 --> 0.961224).  Saving model ...
Epoch: 32 	Training Loss: 0.861942 	Validation Loss: 1.050242
Epoch: 33 	Training Loss: 0.828336 	Validation Loss: 1.131701
Epoch: 34 	Training Loss: 0.817495 	Validation Loss: 1.076523
Epoch: 35 	Training Loss: 0.829824 	Validation Loss: 1.124696
Epoch: 36 	Training Loss: 0.838864 	Validation Loss: 1.062148
Epoch: 37 	Training Loss: 0.840021 	Validation Loss: 1.004353
Epoch: 38 	Training Loss: 0.837052 	Validation Loss: 1.085684
Epoch: 39 	Training Loss: 0.832652 	Validation Loss: 1.035074
Epoch: 40 	Training Loss: 0.814693 	Validation Loss: 1.076445
Epoch: 41 	Training Loss: 0.795244 	Validation Loss: 1.110126
Epoch: 42 	Training Loss: 0.811786 	Validation Loss: 1.044767
Epoch: 43 	Training Loss: 0.804683 	Validation Loss: 1.161828
Epoch: 44 	Training Loss: 0.800759 	Validation Loss: 1.092378
Epoch: 45 	Training Loss: 0.794199 	Validation Loss: 1.090922
Epoch: 46 	Training Loss: 0.809656 	Validation Loss: 1.062465
Epoch: 47 	Training Loss: 0.792386 	Validation Loss: 1.127077
Epoch: 48 	Training Loss: 0.791442 	Validation Loss: 1.034724
Epoch: 49 	Training Loss: 0.778374 	Validation Loss: 0.989902
Epoch: 50 	Training Loss: 0.784115 	Validation Loss: 1.112401
Epoch: 51 	Training Loss: 0.768123 	Validation Loss: 1.044812
Epoch: 52 	Training Loss: 0.764910 	Validation Loss: 1.052471
Epoch: 53 	Training Loss: 0.785023 	Validation Loss: 0.948268
Validation loss decreased (0.961224 --> 0.948268).  Saving model ...
Epoch: 54 	Training Loss: 0.750863 	Validation Loss: 1.128910
Epoch: 55 	Training Loss: 0.749471 	Validation Loss: 1.048802
Epoch: 56 	Training Loss: 0.742631 	Validation Loss: 1.019046
Epoch: 57 	Training Loss: 0.790767 	Validation Loss: 0.962344
Epoch: 58 	Training Loss: 0.771013 	Validation Loss: 1.118378
Epoch: 59 	Training Loss: 0.752799 	Validation Loss: 1.008506
Epoch: 60 	Training Loss: 0.743373 	Validation Loss: 1.098588
Epoch: 61 	Training Loss: 0.760900 	Validation Loss: 1.050495
Epoch: 62 	Training Loss: 0.745241 	Validation Loss: 1.056108
Epoch: 63 	Training Loss: 0.750054 	Validation Loss: 1.128346
Epoch: 64 	Training Loss: 0.732712 	Validation Loss: 1.001609
Epoch: 65 	Training Loss: 0.745477 	Validation Loss: 1.016578
Epoch: 66 	Training Loss: 0.738111 	Validation Loss: 1.037418
Epoch: 67 	Training Loss: 0.734468 	Validation Loss: 1.020058
Epoch: 68 	Training Loss: 0.721990 	Validation Loss: 1.058555
Epoch: 69 	Training Loss: 0.718559 	Validation Loss: 1.022159
Epoch: 70 	Training Loss: 0.711598 	Validation Loss: 1.060030
Epoch: 71 	Training Loss: 0.733466 	Validation Loss: 1.039587
Epoch: 72 	Training Loss: 0.723928 	Validation Loss: 1.011861
Epoch: 73 	Training Loss: 0.726826 	Validation Loss: 1.200244
Epoch: 74 	Training Loss: 0.691118 	Validation Loss: 1.020905
Epoch: 75 	Training Loss: 0.745111 	Validation Loss: 1.037924
Epoch: 76 	Training Loss: 0.699868 	Validation Loss: 1.072613
Epoch: 77 	Training Loss: 0.703926 	Validation Loss: 1.082218
Epoch: 78 	Training Loss: 0.699596 	Validation Loss: 0.993207
Epoch: 79 	Training Loss: 0.700158 	Validation Loss: 1.065973
Epoch: 80 	Training Loss: 0.718534 	Validation Loss: 1.072341
Epoch: 81 	Training Loss: 0.695254 	Validation Loss: 1.030833
Epoch: 82 	Training Loss: 0.681698 	Validation Loss: 0.990697
Epoch: 83 	Training Loss: 0.682091 	Validation Loss: 1.009951
Epoch: 84 	Training Loss: 0.664637 	Validation Loss: 0.964296
Epoch: 85 	Training Loss: 0.700702 	Validation Loss: 0.984724
Epoch: 86 	Training Loss: 0.683968 	Validation Loss: 1.011845
Epoch: 87 	Training Loss: 0.658562 	Validation Loss: 1.001322
Epoch: 88 	Training Loss: 0.702039 	Validation Loss: 0.978508
Epoch: 89 	Training Loss: 0.680450 	Validation Loss: 0.971897
Epoch: 90 	Training Loss: 0.688055 	Validation Loss: 1.063597
Epoch: 91 	Training Loss: 0.678394 	Validation Loss: 0.990956
Epoch: 92 	Training Loss: 0.669466 	Validation Loss: 1.046573
Epoch: 93 	Training Loss: 0.688434 	Validation Loss: 1.064260
Epoch: 94 	Training Loss: 0.673059 	Validation Loss: 1.100582
Epoch: 95 	Training Loss: 0.676138 	Validation Loss: 1.039726
Epoch: 96 	Training Loss: 0.673636 	Validation Loss: 0.973811
Epoch: 97 	Training Loss: 0.705775 	Validation Loss: 1.059109
Epoch: 98 	Training Loss: 0.667903 	Validation Loss: 0.993905
Epoch: 99 	Training Loss: 0.666039 	Validation Loss: 1.074131
Epoch: 100 	Training Loss: 0.653626 	Validation Loss: 1.200771

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 60%.

In [128]:
test(loaders_transfer, model_transfer, criterion_transfer, use_cuda)
Test Loss: 1.003309


Test Accuracy: 72% (604/836)

(IMPLEMENTATION) Predict Dog Breed with the Model

Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan hound, etc) that is predicted by your model.

In [133]:
### TODO: Write a function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.

# list of class names by index, i.e. a name can be accessed like class_names[0]
class_names = [item[4:].replace("_", " ") for item in train_data.classes]

def predict_breed_transfer(img_path):
    # load the image and return the predicted breed
    img = Image.open(img_path)
    img = trfm(img)
    c, h, w = img.size()
    img = img.view((1, c, h, w))
    if use_cuda:
        img = img.cuda()
    output = model_transfer(img)
    _, pred = torch.max(output.data, 1)
    return class_names[pred] # predicted class index

Step 5: Write your Algorithm

Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,

  • if a dog is detected in the image, return the predicted breed.
  • if a human is detected in the image, return the resembling dog breed.
  • if neither is detected in the image, provide output that indicates an error.

You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and human_detector functions developed above. You are required to use your CNN from Step 4 to predict dog breed.

Some sample output for our algorithm is provided below, but feel free to design your own user experience!

Sample Human Output

(IMPLEMENTATION) Write your Algorithm

In [138]:
### TODO: Write your algorithm.
### Feel free to use as many code cells as needed.

def show_img(img_path):
    img = cv2.imread(img_path)
    img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    plt.imshow(img_rgb)
    plt.show()

def run_app(img_path):
    ## handle cases for a human face, dog, and neither
    print()
    if dog_detector(img_path):
        # show image
        print("You are a dog")
        show_img(img_path)
        # return breed
        print("And your breed is - {}".format(predict_breed_transfer(img_path)))
    elif face_detector(img_path):
        # show image
        print("You are human")
        show_img(img_path)
        # return breed
        print("Unfortunately you also look like a - {}".format(predict_breed_transfer(img_path)))
    else:
        print("Woah! The algorithm is grossly wrong, you don't look like anything")
        show_img(img_path)
    print()

Step 6: Test Your Algorithm

In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?

(IMPLEMENTATION) Test Your Algorithm on Sample Images!

Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.

Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.

Answer: (Three possible points for improvement)

The output is better than I expected. I get 72% accuracy which is pretty good, given that the network found the best params around 50 epochs. Points of improvement,

  • For the same human person, the network predicts three different dog breeds. I am not sure if this is ideal behaviour and should certainly be looked into.
  • If there are two dogs in the same picture, the network gets confused. We can certainly use sliding windows at different magnifications to fix this problem.
  • We can have the algorithm add bounding boxes around faces and then we can figure out better where the network is getting things wrong and try and improve there.
In [140]:
## TODO: Execute your algorithm from Step 6 on
## at least 6 images on your computer.
## Feel free to use as many code cells as needed.

## suggested code, below
for file in np.hstack((human_files[15:30], dog_files[:15])):
    run_app(file)
You are a dog
And your breed is - Airedale terrier


You are human
Unfortunately you also look like a - Nova scotia duck tolling retriever


You are human
Unfortunately you also look like a - Nova scotia duck tolling retriever


You are human
Unfortunately you also look like a - Pharaoh hound


You are human
Unfortunately you also look like a - Bichon frise


You are human
Unfortunately you also look like a - American water spaniel


You are human
Unfortunately you also look like a - Brittany


Woah! The algorithm is grossly wrong, you don't look like anything

You are human
Unfortunately you also look like a - Pharaoh hound


You are human
Unfortunately you also look like a - Chihuahua


You are human
Unfortunately you also look like a - Pharaoh hound


You are human
Unfortunately you also look like a - Dogue de bordeaux


You are human
Unfortunately you also look like a - Beagle


You are human
Unfortunately you also look like a - Basenji


You are human
Unfortunately you also look like a - Beagle


You are a dog
And your breed is - Pharaoh hound


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Dogue de bordeaux


You are a dog
And your breed is - Basset hound


You are a dog
And your breed is - Basset hound


You are a dog
And your breed is - Basset hound


You are a dog
And your breed is - Basset hound


You are a dog
And your breed is - Basset hound


You are a dog
And your breed is - Basset hound


You are a dog
And your breed is - Basset hound